All men living in Malmö who were born in 1926-1929 were invited for a health screening examination which included, among many other things, measurement of gamma-glutamyltransferase (GGT). They were followed up and within three years after the screening 963 of 4571 men had been hospitalized. They had 17158 hospital days or almost 6 days per individual each year. The impact of alcohol on the admissions was analysed both according to the International Classifications of Diseases and a new design where all the diagnoses were grouped and coded in conditions which were judged to be alcohol-related, potentially alcohol-influenced and non-alcohol-related. Of the total days in hospital, alcohol psychosis and alcoholism accounted for 13.6 per cent. Altogether 29.2% of the days were caused by alcohol-related and potentially alcohol-influenced conditions. GGT values at the screening investigation were significantly increased in 25% of the hospitalized men. Alcohol-related admissions were seven times as many in men with GGT values in the highest quintile compared with those who had values in the lowest quintile.